Mistral releases its first multimodal model, Pixtral 12B, available on GitHub and Hugging Face, and via API "soon" (Kyle Wiggers/TechCrunch)

Kyle Wiggers / TechCrunch:
Mistral releases its first multimodal model, Pixtral 12B, available on GitHub and Hugging Face, and via API “soon”  —  French AI startup Mistral has released its first model that can process images as well as text.  —  Called Pixtral 12B, the 12-billion-parameter model is roughly 24GB in size.